Edit: For clarification, this question is in regards to silicate use in carbon-based biochemistry, not silica-based biochemistries.

enter image description here

The microorganisms you see above are called diatoms; they are a very diverse group of algae which make silicate cell walls called frustules, which they shape into various forms and arrangements. In other words, they are unicellular plantlike protists that live in glass bodies.

I naturally wondered if this body structure could be applied into a multicellular structure, perhaps forming an analogue to either coral reefs or even woody plants. The main biological constraints I can think of regarding this are:

  • silicates, as inorganic materials, do not grow with an organism, and so would either constrict it until it’s crushed within its own body (as is the fate of diatoms) or the organism would simply outgrow any silicate covering, and need to find/make more.
  • Silicates, as inorganic materials, must be obtained from the environment somehow. For a microorganism, this is easy to do - it can take silicates suspended in water. But for a multicellular organism, it becomes much more challenging - how would such an organism obtain new material to be processed for its outer body?
  • Silicates, as inorganic materials, don’t digest easily. Assuming an organism wants to eat this “glass plant”, then it will need a way to either process the silicates or use them for itself. (This, though, is a comparatively easy concern to address. Niche partitioning is the spice of life~.)

enter image description here

I attempted to create something with these thoughts in mind, and this was the result - a microbial colony mirroring stromatolites, but created with frustules instead of sedimentary deposits. As the autotrophs replicate, new cells leave through a pore at the tip of the frustule; they then cling onto the parent cells, and build their own frustules connected to them, like a microbial coral tower.

While I’m satisfied with this result, though, I’m wondering if this or something similar can be done with a proper multicellular organism. So what circumstances would need to occur to allow for a plant-like or coral-like autotroph to build bodies supported by silicates instead of cellulose or other organic compounds? How effective is this, compared to using organic compounds?

  • $\begingroup$ Phytoliths... $\endgroup$
    – AlexP
    Jan 15 at 6:51
  • 1
    $\begingroup$ It may be hard for a tree to get enough minerals to make a skeleton since the roots only can get so much leachate. Wood almost exclusively comes from CO2 and water. Animals move (or have their food source move via water currents) so it may help them capture more minerals. $\endgroup$ Jan 15 at 22:15
  • $\begingroup$ It seems like "not easy to digest" would be a benefit, since it would disincentivize predation by animals. $\endgroup$
    – Tom
    Jan 16 at 0:23
  • $\begingroup$ @Tom true, but it would also lead to them overpopulating the environment, and that’d cause a ton of overcrowding issues. Even the tallest trees on Earth have insects and smaller organisms that specialize in eating them, to bring their matter back into the environment. $\endgroup$ Jan 16 at 0:55
  • 1
    $\begingroup$ @Jasen I feel the title is concise enough, but I did change the placement of my addendum for the future. Thanks for pointing that out, my apologies. $\endgroup$ Jan 16 at 12:00

4 Answers 4


There are already creatures using silicates for their scaffolding, like the siliceous sponges

The siliceous sponges form a major group of the phylum Porifera, consisting of classes Demospongiae and Hexactinellida. They are characterized by spicules made out of silicon dioxide, unlike calcareous sponges.

Individual siliachoates (silica skeleton scaffolding) can be arranged tightly within the sponginocyte or crosshatched and fused together. Siliceous spicules come in two sizes called megascleres and microscleres.

I guess outside of water the brittleness and low elasticity of silicon dioxide could become too much of an obstacle, but inside water it seems to be pretty fine.

ADDENDUM: The glass sponge: already somewhat shrub like.


There have been several silicon biochemistry questions this month. Why has it suddenly become more interesting? Whatever the reason, I'll skip stuff growing in the sea for now (L.Dutch already covered glass sponges and you already mentioned diatoms) and concentrate only on stuff growing on land.

You may be surprised to learn that there's already a fair amount of silica in vegetation in the world in the form of phytoliths. According to Cycling silicon – the role of accumulation in plants, as much as 47.3 mg per g of dried plant matter of some grasses can be silica, nearly 5%. Silicon as Versatile Player in Plant and Human Biology: Overlooked and Poorly Understood claims as much as 10% of dried mass, but I can't find the reference they're using to back up that figure, possibly because I ran into paywalls.

A quick look at the composition of the human body on Wikipedia suggests that 2.5% of the wet mass of a human is calcium and phosphorous, and given that 53% of the wet mass is, well, wet, some grasses would appear to be as mineralized as you are.

That doesn't necessarily mean there's enough to make something as silicaceous as a tree is woody... lignin (a structural compound in trees and woody shrubs) makes up 17-35% of the dry mass of various kinds of tree.

How effective is this, compared to using organic compounds?

Problem is though that despite how much silicate-based matter there is in the world (silicon and oxygen make up more than 50% of the Earth's crust by mass) it is relatively unreactive.

From The anomaly of silicon in plant biology (PDF):

The soil water, or the "soil solution," contains silicon, mainly as silicic acid, H4SiO4, at 0.1-0.6 mM-concentrations on the order of those of potassium, calcium, and other major plant nutrients, and weli in excess of those of phosphate. Silicon is readily absorbed so that terrestrial plants contain it in appreciable concentrations

(interestingly, this paper also mentions the 10% silica figure, but doesn't source it)

Silicon as Versatile Player in Plant and Human Biology, linked above, says that this silicic acid mostly comes from acidic weathering of silicate rich minerals (not sourced, but does at least sound plausible). Trying to liberate silica directly is very energetically unfavorable, so I don't see any kind of lifeform trying to eat the raw stuff unless it had a fluorine-based chemistry, which is probably a bit too exotic for your needs. Instead, your land corals would need a large enough root system to harvest enough silicic acid from the soil or other water sources, in just the same way that grasses harvest the silicates they need to form their phytoliths.

Your silicaceous corals, then, might need to be somewhere with a good supply of silicic acid, which might imply a climate with more acidic rain that we get falling on igneous rocks or diatomaceous sedimentary rocks. This would ensure high levels of bioavailable silicates, hopefully allowing more substantial silica-based biomineralization than you find on Earth.

To comment on some other bits of your question:

I attempted to create something with these thoughts in mind, and this was the result - a microbial colony mirroring stromatolites, but created with frustules instead of sedimentary deposits.

Either that "coral" is fully covered by the water when the tide is in, or you need some way to transport nutrients to the top (and some reason why it doesn't just widen out below the tidal zone.

Silicates, as inorganic materials, don’t digest easily. Assuming an organism wants to eat this “glass plant”, then it will need a way to either process the silicates or use them for itself

As briefly mentioned above, "digesting" silicate minerals is very energy intensive and I suspect it just won't be done because it isn't worth it. I am now imagining some combination of a goat and a parrotfish, chomping up glassbushes with a beak and regurgitating pellets of sand having extracted the nutritious parts of the plant matter...

  • $\begingroup$ Woah! Such a detailed response! Thank you! So in order to get silica-structured land plants/corals, I’d need environments with regular acid rain. Which prolly means I’d need to tamper with the chemistry of the atmosphere a bit… any recommendations for something I could add that wouldn’t drastically affect the biosphere? The current atmosphere mirrors prehistoric Earth as it was right after the Snowball Event, so there’s plentiful oxygen in the atmosphere and a proper ozone layer formed. The most developed life forms so far are eukaryotic unicellular organisms, so no multicellular life. $\endgroup$ Jan 15 at 11:45
  • $\begingroup$ (cont) As for your other comments, I’d happily share! Regarding my creation, I wouldn’t call that a coral specifically, though I did compare it to one. The closest comparison would be a stromatolite; it’s a structure made by prokaryotic phototrophs, like cyanobacteria, that happen to have converged on the frustule-style body plan. Like stromatolites, they’re fully submerged under high tide, but are exposed in places at low tide. As for why it doesn’t widen out, I figure most of the growth happens upwards, with the prokaryotes crawling upwards to reach the sun, thus creating a tower. $\endgroup$ Jan 15 at 12:00
  • $\begingroup$ (Cont again) I also like that image in your head of the parrot-goatfish! X3 I was imagining something like mastodon or placoderm jaws designed to slice then pulverize the glass, myself. I figured some species might co-opt the plant’s collected silicates into body armor like glassy scales or shells, but that’s more on the speculative deep end. :P $\endgroup$ Jan 15 at 12:04
  • 1
    $\begingroup$ @AtlastheWorldbuilder elevated carbon and sulfur dioxide in the atmosphere might help... volcanism can help there, and provide plenty of fresh silicates exposed to the elements. Having significant extra volcanism over an evolutionarily significant period of time might be awkward, but a shorter window of opprtunity might briefly allow for much larger glass trees that might have existed before. The Deccan traps might have kept going to 30000 years, for example. $\endgroup$ Jan 15 at 12:57
  • $\begingroup$ Hmm... y'kno, I think I might be able to work that in. I do plan on having a mass extinction event from a supercontinent breaking up, similar to the Triassic-Jurassic extinction event caused by the breakup of Pangea. Maybe an increase of volcanism could not only explain the splitting supercontinent, but also allow for copious amounts of acid rain that would give such organisms an opportunity to move their way on land! 0.0! $\endgroup$ Jan 15 at 20:48

Many plants lay down silicon to augment their "skeletons". Horsetails have the most.

No plants I know of need to build their bodies out of silicon lke the glass sponges do. But lots of plants add silicon to their bodies - probably more as a defense than as structural elements. The ability to use biological silica has evolved several times.

Four hundred million years of silica biomineralization in land plants

silica content in various plant groups

Check out Equisetaceae on the left - the horsetails. They are nearly 20% silicon by dry weight.

Horsetails are an ancient group of plants. Back when the earth was theirs some of them got as big as trees. I have looked before for images of the calcium skeletons of horsetails and I finally found some here.

New insight into silica deposition in horsetail (Equisetum arvense)

silica skeletons

Microwave-assisted acid digestion of horsetail, either grown hydroponically in the presence of silicic acid or in plants collected from the wild, resulted in silica deposits and 'skeletons' which were successfully labelled with the fluor PDMPO. Silica was identified in acid digests of all areas of the plant from the rhizome through to spores in the cone. There were no structurally-distinct silica skeletons in the root, only what appeared as diffuse deposits of siliceous materials (Figure ​(Figure1a).1a). Silica skeletons of basal stem showed epidermal-like cells, 30-40 μm wide and 100-300 μm long, with heavily silicified cell walls and approximately equidistant punctate deposits of silica within the walls which were suggestive of the expected locations of plasmodesmata. Each 'silica cell' included an amorphous, spherical silica deposit between 10 and 20 μm in diameter which had the appearance of a nucleus or vesicle. There were also occasional heavily silicified (as indicated by an enhanced fluorescence) skeletons of stomata,

You could riff on plants with silicified tissues and come pretty close to what you proposed in the OP.


"I’m wondering if this or something similar can be done with a proper multicellular organism"

Yes. If the question is limited to only silicates, then the issue is more challenging, although probably still feasible. If the question is broadly, silicon based "organics", then almost definitely.

Starting from a very basic level, most of biology is based on nucleobases of DNA and RNA, with the five nucleobases being adenine (A), cytosine (C), guanine (G), thymine (T), and uracil (U). These are primarily C,H,N,O ring structures. Ex: cytosine is C4H5N3O. Yet, humans already know that Nucleic Acid Analogues and Unnatural Base Pairs already exist, leading to an Expansion of the Genetic Code using lab created nucleobases and amino acids.

With this said, silicon can form many of the same structures and bonds, and even directly substitutes for carbon in many molecules. From the Wikipedia Silicon article "Furthermore, since carbon and silicon are chemical congeners, organosilicon chemistry shows some significant similarities with carbon chemistry, for example in the propensity of such compounds for catenation and forming multiple bonds." (Ex: direct silicon substitution is often viewed as an issue in graphene).

In terms of Cyclosilicate ring structures, an example with very similar organic character is: Hexamethylcyclotrisilazane which is a ring structure of C6H21N3Si3 (Note the Si is actually the ring and the carbon are just edge attachments).

Silicon can also form multi ring structures, such as some of the Inosilicates like tremolite or pellyite. It can also form long graphene like sheet structures of rings such as the Phyllosilicates. Apparently, chemists have only recently started realizing though that silicon can form many of the exact same aromatic ring structures (2010 paper)

The implication of this being, that alternatives to the basic building blocks can "probably" be synthesized using C => Si substitutions. (Note: this is unproven personal supposition)

This also leads into the other idea, that Organosilicon Chemistry is already a concept. Silicone is perhaps one of the most well known products of this line of research, and an excellent example of how not all silicon compounds are stereo-typically "glassy." When it bonds in a chain of the general form (⋯−Si−O−Si−O−Si−O−⋯) it becomes that soft rubbery material use to seal windows against air and water.

Silicon based life, or silicon-carbon mixture life could make all kinds of uses for a material that's relatively pliable, resistant to chemical attack, and fairly easily replaceable with new additions.

Silicon is also already used for Water Glass (sodium silicate) and Silica Gel. Silica gel especially can be useful for silica-organic analogues because it can adsorb ~40% of its weight in water, and then readily release the water by heating at 120°C and return back to its starting state.

Finally, large scale structural materials of silicon, or silicon-carbon (such as Silicon Carbide) are both quite viable, in addition to the Silicon Nitrides. Silicon Carbide and Silicon Nitride are both thermodynamically stable, extremely tough and durable, and can be readily shaped into numerous different forms.

Finally, several creatures already use silicon in the formation of shells and exoskeletons. Diatoms and sponges have already been mentioned that cause an undersaturated solution of silicic acid to polymerise to form silica using an enzyme process that is not especially well understood. The Radiolaria groups is another category that also uses silica.

So at least in a very theoretical framework: 1) The basic building blocks of DNA / RNA appear to show at least the possibility of a silicon substitution for alternatives. 2) Hard structural framework materials (skeletons/endoskeletons, exoskeletons) appear to be fabricatable using either silicon alone, or a silicon-carbon mixture. 3) Soft, flexible gel-like structures similar to skin appear to be fabricatable such as Silicone, Silica Gel, or Water Glass, which also have the possibility for water and nutrient regulation at close to Earth temperatures, and might work as substitutes for cell walls. 4) Silicon is also being investigated for possible uses as muscles, so the concept is at least there for large scale actuators.

  • $\begingroup$ Thanks for the reply! I’ll admit, though, that while the information is appreciated, the question is less about silica-based biochem specifically, and more about silica use in carbon-based biochem. The project is already a bit too far developed to drastically rewrite such things as DNA alternatives or such; however, the idea of carbon-based life forms able to adapt and use silica in their biology is an interesting concept that I’m considering exploring further! In any case, thanks for the extra information! This is helpful to me all the same <3 $\endgroup$ Jan 15 at 21:52
  • 1
    $\begingroup$ You're welcome. Was mostly trying to answer the question from the "ground up" and at least one of the questions would be "could the equivalent of DNA form." Little confused, since you say "more about silica use in carbon-based biochem" and then "carbon-based life forms able to adapt and use silica". However, silicone and silica gel would both be possibilities in that that type of system and there's some work mixing silicone-cellulose or adding silicone to cellulose $\endgroup$
    – G. Putnam
    Jan 15 at 22:07
  • $\begingroup$ Ooo! Today I Learned! :3 $\endgroup$ Jan 15 at 22:14

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .